Skip to content
Lizely

Jewel Match

Solve a fully disclosed match-three cascade with mouse, touch, or keyboard controls and a unique three-swap minimum independently verified by breadth-first search.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Select one gem, then select an orthogonally adjacent gem; the exchange is accepted only if it immediately creates a horizontal or vertical line of at least three identical gems.
  2. 2.Let every matched line clear, then watch surviving gems fall and the disclosed Sapphire-Ruby-Topaz-Emerald queue refill each column until all cascades are stable.
  3. 3.Clear 12 gems to earn exactly 1,000 points; the independently verified unique minimum route uses three legal swaps, while two illegal swaps require Restart.

About Jewel Match

Jewel Match presents match-three as a small deterministic logic level instead of an endless random score chase. The board is a 4 by 4 grid containing ruby, sapphire, emerald, and topaz gems. Each gem type appears exactly four times at the start, and the opening board has no existing line. The objective is to clear 12 gems. A legal route reaches that target in three player swaps, while an independently written breadth-first oracle proves that no route can finish in fewer and that the minimum route is unique.

The interaction follows the cited match-three rules. Select one gem and then a gem sharing an edge with it. Diagonal cells are not adjacent. The two gems exchange positions only when that exchange immediately creates a horizontal or vertical straight line containing at least three identical gems. A swap that creates no qualifying line is rejected and the original board is restored. PySolFC's maintained Match Three documentation describes adjacent, match-forming exchanges, clear, gravity, refill, and repeated cascades. PuzzleBento independently confirms neighboring selection, horizontal or vertical 3-plus lines, rejected swaps, falling columns, new pieces, and chain reactions.

The fixed board, target, refill order, score, controls, and two-error rule are Lizely product choices. Read row by row from the top, the initial board is R S R S, E T R T, E S E R, and T T E S. The letters mean Ruby, Sapphire, Emerald, and Topaz. Every cell exposes its row, column, and full gem name to assistive technology, and the letter remains visible so no decision relies on color perception alone.

After a legal swap, all qualifying horizontal and vertical lines are collected at the same time. Overlapping lines clear their union once rather than double-counting an intersection. Within each affected column, surviving gems keep their relative order and fall toward the bottom. Empty positions at the top are then filled from one disclosed queue. If that stable refill forms another line, the clear, gravity, and refill loop repeats automatically before the next player input. Each loop is one cascade wave.

The refill queue is not random. It contains 64 gems in a repeated Sapphire, Ruby, Topaz, Emerald cycle, giving 16 of each type. Columns are processed from left to right. Within a column, the earliest queue item falls into the lowest vacancy, followed by later items above it. This convention completely determines every post-clear board. The fixed objective consumes only the first 12 queue entries, leaving a large defensive reserve and avoiding network, clock, seed, or browser-dependent behavior.

The unique shortest route is fully disclosed. First exchange indexes 10 and 11, the adjacent cells at row 3 columns 3 and 4. This forms a vertical ruby line at indexes 2, 6, and 10 and clears three gems. Next exchange indexes 11 and 15, forming a horizontal sapphire line at indexes 9, 10, and 11 and clearing three more. Finally exchange indexes 5 and 6. A vertical line clears indexes 5, 9, and 13; gravity and the fixed refill then form a second line at indexes 13, 14, and 15. That two-wave cascade clears six, bringing the total to 12.

Touch and mouse controls use the same two-click model. The selected source receives a visible ring and an accessible pressed state. Choosing the same cell again cancels selection without penalty. Choosing a non-neighbor or a neighbor that creates no line is an illegal swap. No gem, refill cursor, clear total, or legal move count changes after that attempt. The feedback explains whether adjacency or the missing match caused the rollback.

Keyboard play covers every action. Arrow keys or W, A, S, and D move the logical cursor around the 4 by 4 grid, wrapping at each edge. Enter or Space selects the cursor gem and then attempts the swap with a chosen neighbor. R restarts the level. The shared double-Escape boss key replaces the board with a neutral spreadsheet view; repeating the gesture returns to the unchanged game. Repeated keydown events are ignored so holding a key cannot accidentally make two selections or spend two errors.

One illegal swap is repairable. The player can immediately choose the unique valid opening exchange and continue from the same board. A second illegal swap deadlocks the run until Restart. This transparent threshold is not claimed as a historical match-three rule. It gives browser acceptance tests a deterministic failure path and makes rejected input visible. Restart restores all 16 initial gems, refill cursor zero, clear total zero, move count zero, cursor position zero, no selection, no mistakes, and no terminal state.

Clearing 12 gems awards exactly 1,000 points. The score is fixed: it does not multiply cascades, reward speed, or penalize a longer legal route. Once complete, the board freezes. Cursor movement, selection, and swaps cannot change gems, counters, refill state, or score. Deadlock is also frozen. This makes the terminal result auditable and prevents a delayed keyboard event from changing a captured completion state.

The release tests do not use the production matcher as its own proof. An independent oracle represents the same gems as A, B, C, and D, scans rows and columns with separate loops, implements its own gravity and refill routine, enumerates all orthogonally adjacent pairs, and searches legal states by depth. It finds exactly one legal opening exchange and exactly one shortest path to 12 cleared gems at depth three. The product solver is then compared against that independently authored route.

Golden tests also isolate the rule boundaries. They prove that an edge neighbor is accepted while a diagonal is rejected, identify concrete horizontal and vertical triples, verify a no-match rollback, check the first resolved board and its gem inventory, and show that the first swap consumes exactly three refill gems. They confirm the three move clear totals of 3, 3, and 6, cascade wave counts of 1, 1, and 2, and a refill index equal to the total number cleared.

Refill exhaustion is handled atomically even though the published queue is much longer than the level requires. A resolver given an empty queue returns failure instead of exposing a half-cleared board. Runtime converts any such defensive failure into the same unchanged-board error path. Tests also verify that every normal resolved board contains 16 gems and that illegal input preserves the serialized board and queue position.

Everything runs locally in the browser. No board, score, move history, or interaction is uploaded. No account, paid API, random service, or new runtime dependency is required. The grid uses native buttons with touch-friendly dimensions, visible focus treatment, English labels, and compact responsive spacing for a 390-pixel viewport. The bounded 4 by 4 search and direct move resolution remain small enough for immediate interaction.

Jewel Match is recreational entertainment. It is not an intelligence test, aptitude assessment, classroom placement tool, psychological instrument, hiring screen, or health measure. Knowing the disclosed route makes later runs trivial, and a key slip can count as an error. For an unaided attempt, inspect the lower-right corner first: the only opening exchange must move one emerald away so three rubies align vertically.

Methodology & sources

Encode the fixed 4 by 4 board in row-major order and reject every exchange except orthogonally adjacent pairs whose swapped board immediately contains a horizontal or vertical run of at least three. Clear the union of all matched indexes, preserve survivor order under column gravity, and refill columns left to right from a fixed 64-gem Sapphire-Ruby-Topaz-Emerald cycle, repeating until stable. Breadth-first search over board, refill index, and cumulative clear count finds the first goal at three swaps. An independent letter-token oracle separately implements scanning, gravity, refill, legal-pair enumeration, and depth traversal, proving a unique route of 10-11, 11-15, 5-6 with clear totals 3, 3, and 6. Runtime invalid swaps preserve board and queue, the first remains repairable, the second deadlocks, completion awards exactly 1,000, and terminal states freeze until Restart.

Frequently asked questions

Why did my jewel swap roll back?
The cells were not orthogonal neighbors, or the exchanged board created no immediate horizontal or vertical line of three or more. An illegal attempt changes no gem or refill position.
How are cascades resolved?
All current 3-plus lines clear together, survivors fall to the bottom of each column, and a fixed queue refills the vacancies. New matches repeat the loop before the next input.
What is the shortest solution?
Independent breadth-first search proves one unique three-swap minimum: indexes 10-11, then 11-15, then 5-6. Those moves clear 3, 3, and 6 gems.
Does the 1,000-point result measure ability?
No. The fixed score, disclosed board, refill queue, and two-error stop are game behavior. This level is entertainment, not a standardized intelligence, education, employment, health, or psychological assessment.

Related tools

Mini Games guides

View all